[17840] in Athena Bugs
Re: sun4 8.3.27: discuss
daemon@ATHENA.MIT.EDU (Kev)
Mon May 15 10:13:48 2000
Message-Id: <200005151413.KAA01008@mint-square.mit.edu>
To: Brad Thompson <yak@MIT.EDU>
cc: bugs@MIT.EDU
In-Reply-To: Your message of "Sat, 13 May 2000 11:27:19 EDT."
<200005131527.LAA02307@manatee.mit.edu>
Date: Mon, 15 May 2000 10:13:44 -0400
From: Kev <klmitch@MIT.EDU>
> Occasionally, discuss has the following problem:
> 1) I type "ckm". It says that some meeting has been changed.
> 2) I type "nm". I then type "next" to read the new transaction.
> 3) I type "next". It says I have no more unread transactions.
> 4) I type "ckm" again. It says that the same meeting has been
> changed.
>
> I can repeat this ckm-nm-next cycle indefinitely. Inspection of the
> .meetings file shows that discuss is not updating the last-read-article-numbe
> r
> field. This bug is reproducible. To see it for yourself, launch
> discuss and add the meeting menelaus.mit.edu:/usr/spool/discuss/netusers.
> Show transaction 618, then "next" to 619. Transaction 619 will not get
> marked read unless you modify the .meetings file directly. I have
> checked this with an empty .meetings file under the sipbtest account,
> so I am convinced that it is not an account specific thing.
I have seen this behavior in connection with long messages only, with less...
it may happen with more, but I haven't checked that. What seems to be
happening is that less is executed with the file on stdin; if you do not
page to the end of the file before hitting 'q', less returns non-zero, and
discuss assumes that there was a failure in reading the message. Curiously,
you can then type 'next' again, and it will retrieve the next message in the
meeting, of course setting your last read transaction to the correct thing.
--
Kevin L. Mitchell <klmitch@mit.edu>
